How much does switching away from Intercom actually cost?
Switching away from Intercom costs more than the new vendor's invoice. The visible part is whatever number forces the decision: an Advanced-plan renewal at $85 a seat instead of Essential's $29, or a Fin AI bill that outgrew the support team's own headcount cost. Three costs nobody puts in that comparison don't show up until after the decision is made: the contract you're still inside of, which auto-renews annually and is non-refundable outside a narrow set of exceptions; the conversation history, which Intercom will only hand back in capped, partial batches; and the rebuild, every macro, bot flow, and integration recreated from scratch on whatever comes next.
That's not an argument against switching. It just means the honest total is closer to a quarter's worth of the old bill plus a real chunk of someone's time, not the sticker-price gap between two pricing pages.
What triggers the decision
Intercom prices seats in three tiers: Essential at $29 per seat a month, Advanced at $85, and Expert at $132, per Intercom's own pricing page. Fin, the AI agent, bills separately at $0.99 per resolved conversation. Teams rarely leave over Essential's price. They leave when growth pushes them into Advanced for a feature they now need (custom reports, workflows, more than a couple of inboxes), and the per-seat number multiplies against a support team that also grew.
The complaint shows up reliably outside Intercom's own marketing. On Hacker News, one founder described a custom Intercom plan jumping from $300 a month to $2,000 with no material change in usage, and being charged $1,800 even after repeatedly asking to stay on a $100-a-month Chat plan instead; the founder says Intercom only refunded the difference after "a stressful evening" of pushing back, in a widely-cited 2019 thread. A separate commenter, describing why two different companies he'd worked at had both left Intercom, put the underlying mismatch plainly: "their pricing does not make sense for B2C". Per-seat and per-resolution pricing both scale with something other than the business's own margin: seats scale with headcount, resolutions scale with conversation volume, and neither one tracks whether the business behind them is a $200-a-month SaaS seat or a few-dollar consumer order. Our companion guide on why Fin's per-resolution pricing gets expensive covers that second trigger in more depth; this one is about what happens after the decision to leave gets made.
What the switch cost Coastal Crate
Coastal Crate ships a monthly subscription box of regional snack brands, and clears a few dollars of margin per box after fulfillment. Desmond Achterberg, who runs support for a five-person team, had been on Intercom's Advanced plan for the custom reporting his ops lead needed. A referral-program launch tripled inbound volume for six weeks, and Coastal Crate added three seasonal seats to keep response times reasonable. Three years of Intercom tags included a "flavor request" label the team used informally to track which regional snack flavors kept coming up in customer messages, the closest thing they had to a running count of what customers actually wanted next.
The renewal invoice was the trigger:
Desmond, in the team's weekly ops sync: "We're paying for eight Advanced seats now, three of them seasonal, and the contract auto-renews in five weeks whether we keep them or not. If we're dropping back to five seats after the holiday rush anyway, this is the week to actually decide if we're staying on Intercom at all."
The decision to leave took an afternoon. The actual migration took six weeks, and the buckets below are where that time went.
issue filed·
report writtenThe contract you're still inside of
Intercom's terms auto-renew annual subscriptions for successive twelve-month periods unless either side gives written notice at least 30 days before the term ends, and fees are non-refundable outside a narrow set of exceptions, per Intercom's terms and policies. Practically, that means the exit date isn't the day you finish setting up a new tool. It's whichever comes first: the 30-day notice window before renewal, or eating another full year on a plan you've already decided to leave. Coastal Crate's renewal fell mid-migration, and the team sent the non-renewal notice before the export was finished, accepting a few weeks of running both systems in parallel rather than risk a second annual term.
The data you're trying to take with you
Intercom's own export tooling is built for reporting, not for a full-fidelity migration, and the gaps matter for anyone leaving. CSV exports carry conversation metadata, timestamps, tags, and IDs, but not the actual message content. Bulk exports through the workspace UI do carry message content, but only as PDF or text files, capped at 200 to 300 conversations per operation, per Intercom's documentation. Getting that content back as structured JSON instead means the Conversations API, pulled one conversation at a time, or a continuous export to Amazon S3 or Google Cloud Storage that an engineer has to configure, writing data in batches of 200 conversations per file. Intercom states its CSV and JSON exports are generally available for up to two years, per the same article, and confirms the same two-year ceiling for the cloud storage historical export specifically, per Intercom's cloud storage documentation; it doesn't spell out a separate retention window for the PDF/text bulk export, so treat two years as the floor on that path, not a guarantee. A support team that's been on Intercom for four years is migrating at most two years of history through any of these routes and losing the rest, unless someone already pulled it out earlier.
For Coastal Crate, with roughly three years of conversation history and no engineer to spare for a cloud storage integration, this meant a support ops contractor spending four days running Intercom's bulk UI export in 250-conversation PDF batches, then a fifth day confirming nothing had silently dropped.
The workflows you have to rebuild from nothing
The invoice comparison never includes this part, and it's usually the largest one. Every macro, saved reply, routing rule, and bot flow built inside Intercom over however many years lives in Intercom's format and doesn't transfer. Whatever replaces it needs its own triggers, its own canned responses, its own help center articles, and its own bot or AI-resolution equivalent rebuilt from whatever documentation exists of the old setup, on top of the migration labor, not instead of it. For Coastal Crate that meant Zendesk, whose Suite Team plan runs $55 per agent a month billed annually, per Zendesk's pricing page, which for the team's steady-state five seats is its own new $3,300-a-year line item before a single macro gets rebuilt. Coastal Crate's five macros and two routing rules were small enough to rebuild in a day. A team with two years of accumulated automation is looking at weeks, and the honest advice is to budget for that before signing anything, not after.
Where the switch stops being just a helpdesk problem
The three costs above don't touch the thing that made Coastal Crate's "flavor request" tag worth keeping in the first place: a running sense of which flavors kept coming up, and from how many separate customers, without someone re-reading three years of tickets to reconstruct it. That tag lived entirely inside Intercom's own tagging, which means it resets at the exact moment the export caps its history at two years and whatever survives gets re-tagged, by hand, in Zendesk's version of a tag. A count that took three years to build doesn't carry over a helpdesk swap unless something outside both helpdesks was keeping it independently.
That's the specific, narrow gap Modem sits in front of, not the contract, not the export, not the rebuild, just the counting. It reads Intercom and whatever replaces it, Zendesk included, the same way: conversations, replies, and CSAT as they happen, classified into topics, per Modem's Intercom integration. A Coastal Crate-style flavor-request topic built from months of Intercom conversations keeps counting the same request once the identical conversations start arriving from Zendesk instead, because the classification lives in Modem rather than in either helpdesk's tagging. Modem also prices per organization rather than per seat, unlimited users on every plan with usage-based overages beyond included volume, per Modem's pricing page, so the number that triggered the migration in the first place doesn't reappear on the other side of it.
This guide is written by the company that makes that product. It has nothing to say about the contract exit, the export caps, or the rebuild hours above, those land the same whether or not anyone buys Modem, and it doesn't answer a single support ticket on its own. Read the pitch with that in mind.
If per-resolution AI pricing is the specific number behind your decision, our breakdown of Fin's cost per resolution walks through when that math flips against a business. For a wider read on what a dedicated tool adds beyond either helpdesk's native tagging, see the best tools for mining feedback out of Intercom conversations.
The honest total
Coastal Crate's real migration cost was the five days of export labor, roughly a day of workflow rebuilding, the new platform's own $3,300 annual seat cost, and five weeks of running two support tools in parallel while the non-renewal notice cleared. None of that showed up on the pricing-page comparison that started the conversation. If a seat or resolution bill is what's pushing the decision, run those four numbers, contract exit timing, export scope, rebuild time, and the new platform's own cost, before assuming the switch pays for itself in the first quarter. Often it still does. It's just a different number than the one on the invoice.
